1. Field of the Invention
Embodiments of the present invention relate to utilizing multiple graphics subsystems in a computer system.
2. Related Art
Computers often include multiple graphics processing units (GPUs). Often, a computer will supply both an integrated GPU (IGPU) and a discrete GPU (dGPU). Typically, an IGPU is integrated into a portion of a processor chipset, e.g., the north bridge. IGPU's offer a low-cost, low-power video solution. However, many users need a more powerful graphics option, and will use a dGPU to get the functionality they require.
Even for users with high-end graphics requirements, however, it may be advantageous to make use of the IGPU in some scenarios. For example, as the IGPU typically draws much less power than the dGPU, switching to the IGPU for non-graphics intensive work may offer an advantage, in terms of power consumption and/or battery life.
One approach to utilizing both the IGPU and dGPU involves a backend switching mechanism, typically, such as a switching mechanism that allows for selecting whether a display is driven by the IGPU or dGPU. However, this approach involves several issues. Aside from the expense of the actual switching mechanism, the user might experience some disruption in their display, particularly with flat-panel displays, as the low-voltage differential signaling (LVDS) utilized with such displays does not handle such physical switching gracefully.
A second approach involves a making a selection between the frame buffers (and the data contained therein) associated with both the IGPU and dGPU. While this approach eliminates many of the concerns inherent in the previous approach, it introduces several new problems. For example, this approach relies heavily on copying data between frame buffers. This duplication results in performance bottlenecks, e.g., across the PCI-E bus, and also consumes additional bandwidth in the system memory bus.